For output, either in teraconverter or paraconverter mode, is it possible to force an exact tile-size based on height/width/depth, even if there are some small edge tiles? The default seems to load balance the output into tiles that are approximately the same size, using the input height/width/depth as rough guidelines.
My use case is a secondary step to create a precomputed cloud-volume -- or any format viewable by neuroglancer. My current understanding is that the parallel uploading of such a volume would benefit from exact tiling, preferably with dimensions in powers of two.
chrisroat
commented
4 years ago
For output, either in teraconverter or paraconverter mode, is it possible to force an exact tile-size based on height/width/depth, even if there are some small edge tiles? The default seems to load balance the output into tiles that are approximately the same size, using the input height/width/depth as rough guidelines.
My use case is a secondary step to create a precomputed cloud-volume -- or any format viewable by neuroglancer. My current understanding is that the parallel uploading of such a volume would benefit from exact tiling, preferably with dimensions in powers of two.